PSA screening has contributed to a dramatic increase in the number of prostate cancer cases diagnosed and a subtle decrease in mortality from this disease. Unfortunately, many patients still present with aggressive, potentially lethal disease either because 1) the patient did not participate in a screening program, 2) the program did not identify the disease early enough for cure or 3) the treatment chosen did not eradicate the disease. At the same time, many men go through years of unnecessary screening and/or are diagnosed with prostate carcinoma with low metastatic potential. What is needed are markers to identify men, not at high risk of having prostate carcinoma, but at high risk of dying of prostate carcinoma. Identification of high risk men prior to developing the disease would allow intense screening and/or prophylaxis while sparing those at low risk years of screening. Identification of high risk men after diagnoses would allow targeted therapy while sparing those at low risk potentially morbid treatment. We believe that analysis of common polymorphic variants will allow exactly this risk stratification. The objective of this application is to identify which DNA polymorphisms that are associated with risk for metastatic prostate carcinoma and therefore are potential markers for aggressive prostate cancer. We propose to perform a case-control study to identify polymorphic variants associated with advanced prostate cancer using patients with metastatic prostate carcinoma as our cases and individuals with little or no risk of ever developing prostate carcinoma as our controls. There are two specific aims to this proposal. (I) To establish a bank of somatic DNAs from advanced prostate cancer patients and race matched controls. (II)To search for associations between genotypes and risk of advanced prostate carcinoma. The rationale for studying patients with metastatic disease is two fold. First, patients with metastatic disease unequivocally have clinically important prostate carcinoma. As a result, there are no patients with indolent disease to obscure significant associations. Second, since these are the patients in whom earlier identification has the potential to make the greatest impact, they are the patients who should be studied. At the conclusion of this study, we will have a panel of markers with the potential to predict, in future prospective studies, which patients are at high risk of eventual death from prostate carcinoma.